Machine learning for medicine: towards efficient, optimized healthcare delivery

Machine learning has vast potential for disrupting healthcare. In this talk I will discuss various work revolving around innovations in machine learning algorithms and systems for tackling problems such as asthma risk prediction, cost reduction in hospitals, and identification of high-risk patients. Innovative technologies include phenotyping algorithms based on higher order tensor factorization, web services compatible with standardized data models such as HL7 FHIR, and predictive modeling pipelines allowing for physicians and clinical researchers with limited programming experience to construct and test models on the fly.